Background: Pancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer-related death with a 5-year survival rate of 10%. Novel strategies for advanced PDAC are a critical need. A DNA hypomethylating agent (HMA) increased tumor-infiltrating effector T cells and significantly prolonged survival in combination with immune checkpoint blockade (ICB) in a PDAC mouse model (Gonda et al 2020). We hypothesized that combining HMA with ICB will lead to therapeutic benefit in pts with advanced PDAC. Methods: This is an open-label, single-arm, single-center, Phase 2 trial of AZA plus pembro in pts with unresectable or metastatic PDAC. Pts were treated with AZA 50 mg/m2 subcutaneous daily for 5 days Q4W beginning week 1 day 1 followed by pembro 200 mg IV Q3W starting week 3 day 1. Key eligibility criteria included documented progression on or following first-line systemic treatment, ECOG PS 0-1, and adequate organ function. The primary objective was progression-free survival (PFS). Thirty-one evaluable pts were required to detect an improvement in PFS from 2 mo to 4 mo with a one-sided p-value of 0.05 and 80% power. Secondary endpoints included safety and tolerability, overall response rate (ORR), disease control rate (DCR), and overall survival (OS). Data cutoff was February 10, 2022. Results: Between Oct 2017 and Sept 2021, 36 pts were enrolled (median age 62.5, 75% white, 72% male). At data cutoff, 34 and 31 pts received at least 1 dose of AZA and pembro, respectively. Median PFS for ≥1 dose of AZA was 1.48 mo (95% CI: 1.35, 1.74) and for ≥1 dose of pembro was 1.51 mo (95% CI, 1.38, 3.42). The median OS was 4.67 months. Among the 34 pts, 3 (8.8%) experienced a partial response (PR) by RECIST with a DCR of 32.4%. Of the 3 pts with PR, 2 received 35 doses of pembro and continued beyond 2 years with an ongoing PR. One pt remains on pembro 20 months after study completion. None of the pts with a PR were microsatellite instability-high or tumor mutation burden high, and next-generation sequencing of the 3 cancers identified a BRCA1 variant and POLE variant. Treatment-related AEs (TRAEs) occurred in 20/34 pts (59%), the most common of which were diarrhea and fatigue. Grade ≥3 occurred in 7/34 (21%) including 1 immune-related grade 5 event (encephalitis). Conclusions: AZA plus pembro demonstrated a tolerable safety profile but no PFS benefit compared to historical controls. Our data suggests combined epigenetic therapy and ICB may expand therapeutic options in a subset of pts with PDAC. Further investigation is needed to identify biomarkers to predict response and elucidate effective timing, sequencing, and combination of epigenetic agents. Follow-up for OS is ongoing. Correlative analysis of epigenetic effects and characterization of the tumor immune microenvironment will be reported. Clinical trial information: NCT03264404.